International Symposium and Seminar on The Safety of High Speed Craft, February 1997, London, UK

DYNAMIC STABILITY OF PLANING SHIPS

S L Toxopeus, Delft University of Technology, The Netherlands
J A Keuning, Delft University of Technology, The Netherlands
J P Hooft, MARIN, Wageningen, The Netherlands

SUMMARY
At present, most of the dynamic research on planing ships has been directed towards analysing the ship's motions in either
the 3-DOF (Degrees Of Freedom) mode in the longitudinal vertical plane or in the 3-DOF or 4DOF mode in the lateral
vertical plane.
For this reason Delft University of Technology and MARIN have started the set-up of describing the dynamic behaviour
of planing ships in a 6DOF mathematical model. This research program consisted first of all in developing a 6-00F
computer simulation program in the' time domain. Such a simulation program is to be used to predict the response of
these type of vessels to disturbances during high speed sailing.
For describing the behaviour of planing ships in still water static tests have been executed with two planing hull forms in
the towing tank of Delft University of Technology. The test program consisted of measuring three force- and three moment
components as a function of the pitch, rise (draught), roll, drift and speed of the model.
At a next stage a model test program is anticipated to determine the added mass and damping components of these two
hull forms and also the rudder forces. In the meantime the program is in operation while using empirically estimated values
for these quantities.
In this paper the set-up of the mathematical model will be presented. Also a discussion will be given about the use of
these static contributions in a time domain simulation to model the behaviour of the ship.

AUTHORS' BIOGRAPHIES
Mr S L Toxopeus graduated from the Oelft University of
Technology in 1996. He is currently employed as project
manager at the Maritime Research Institute of the
Netherlands in Wageningen.
Or J A Keuning graduated from the Oelft University of
Technology in 1977 and obtained his doctorate at Oelft
University of Technology in 1994. He is employed as a
lecturer at the Ship Hydromechanics Department of Delft
Univsrslty jot Technology, his main subject being the
dynamics of advanced naval vehicles, including sailing
yachts.
Or J P Hoof! graduated from the Delft University of
Technology in 1962 and obtained his doctorate at Delft
University of Technology in 1970. He has been employed
at MARIN since 1962. He is currently senior project
manager involved in developing mathematical models for
simulating the behaviour of various kinds of marine
structures.

1.

INTRODUCTION

The motions of planing craft have been the subject of

many research projects during the last few decades. The
dynamic research was largely directed towards analysing
the motions of the ship in either the longitudinal vertical
plane for three degrees of freedom, see e.g. [13J, [19J, or
in the four-OOF mode in the lateral vertical plane or
horizontal plane [9], [14]. Reviewing the literature about
dynamic stability of high speed craft, it appears that a

1997: The Royal Institution of Naval Architects

mathematical model with six degrees of freedom does not

exist.
At present it is thought that incorporating all six degrees
of freedom into the mathematical models becomes
increasingly important. Instabilities have been reported in
both longitudinal and lateral directions with motions ranging from rapid loss of running trim, progressive heeling,
broaching or a sudden combined roll-yaw motion, possibly
resulting in crew injury or craft loss (Refs. 14, 20 and 21).
Most instabilities are suspected to originate from coupling
between the six degrees of motions. For example, large
bow-down trim angles will most likely result in transverse
instability: yaw motions. For reliable prediction of the
dynamic stability and manoeuvrability of planing craft, all
six degrees of freedom have to be accounted for.
For describing the behaviour of planing ships in still water,
static captive model tests have been executed with two
planing hull forms in the towing tank at Delft University of
Technology, see Reference [17]. The test program
consisted of measuring three force and three moment
components as a function of the pitch, rise (draught), roll,
drift and speed of the models. At a future stage a model
test program is anticipated to determine the added mass
and damping forces of these planing hull forms and also
the rudder and propeller forces acting on the models. At
present, the computer program is in operation while using
empirically estimated values for these quantities.
In this paper, the set-up of the preliminary non-linear
mathematical model for six degrees of freedom based on

Rotations in both coordinate systems are positive if

clockwise. looking in positive direction. The roll angle Ill.
the pitch angle
and the yaw angle 'l' are rotations
around the ship-fixed x, y and z axis respectively. The drift

the data obtained from the model experiments will be

described. This mathematical model has been incorporated in a time-domain computer simulation program in
order to predict the dynamic stability and manoeuvrability
I"'\f
!:I f"\'~"i,"., f:!:htf"'ll
.......... ,..,'lio4t UII~ ~'IIt-'J

~l'!.J.""l D~r!r 4'l"'I".I"'\P'\~"'"

~vg

\-'::;;11Ii:;;:1'1i:t~

r1l::1 T ....
''''''IQ" l'
vJ- 11'C:l'

1"l.

""'J"".llf+... AI

4'

angleis usedto definethe non-dimensional lateralvelocity

'YWUU,.;) '-"I ~VliltJ

simulations performed will be included and discussed.

Also, recommendations will be made for further study, to
increase the accuracy of the mathematical model.

component:

2.

with

COORDINATE SYSTEM

The coordinate systems used in this study are cartesian

coordinate systems. One coordinate system is the
ship-fixed coordinate system, with the x-axis pointing
forward perpendicular to the baseline of the ship and the
z-axis downward. The y-axis is pointed to starboard. The
origin is at the intersection of frame 0 and the baseline of
the model.

u being

the longitudinal velocity component.

3.

M0DEL TESTS

3.1

MODEL PARTICULARS

The models used for this study are Model 233 and Model
277 of Delft University of Technology. Model 277 is based
on the Clement and Blount [1] 62 series with a deadrise
of 25 degrees. Keuning [6] performed seakeeping tests
with this model. Mode! 233 is used by Keuning et al [8]
during experiments with models with warped bottoms.
The main particulars are stated in the table below:

The x-y-plane of the earth-fixed coordinate system

coincides with the undisturbed water surface, the z-axis is
pointed downward.

Symbol
Type
Length

Model 233

Model 277

Deadrise 25

Twisted bottom

1.50m

1.50m

e.;

0.367m

O.367m

Ap

0.450m 2

0.4589rrr

Centre of planing area forward of ord 0

CAP

48.8%L

48.8%L

Length/Beam ratio

UB

4.09

4.09

Max. beam at chine

Projected area

Mass model ine!. transducer

9.45kg

6. 67kg

Longitudinal centre of reference

LCOR

O.726m

O.726m

Vertical centre of reference

VCOR

O.080m

O.OBOm

In Figure 1 the body plans of both models are included

1\

t1

~
~
. .
. .

It-

k2L
~
/7 t7'7"--7

I
I~\~ ~
~

r-, r-,

7i"'r?

[( fl/

7 /7

~~

--j

. .

Cl

.L%
. .
,

VIJ

-j

Figure 1 Body plans of Model 233 (above) and Model 277 (below)

77 i7-?-

l7

3.2

3.3

TEST PROGRAM

In the present study some static captive tests have been

performed with both models during which the forces and
moments in the six degrees of freedom were measured.
These tests will only provide a limited amount of
information about the hydrodynamics of planing ships.
Therefore in further studies additional tests will be
performed to determine the remaining hydrodynamic
characteristics.

EXPERIMENTAL SET-UP

Two six-component transducers were fixed into the

models, evenly spaced around the centre of reference.
Adding the components of the transducers. three forces
and three moments about the centre of reference could be
found.
The measurement of the forces was divided in two parts:
a velocity independent part at speed U _ oms" and a
velocity dependent part, obtained by measuring the
change in forces due to the towing speed. Adding the two
components yields the total force acting on the hull during
the run.

The following variables have been tested:

Speed U:
the speed has great influence on the position of the ship
relative to the free water surface. A change in speed will
result in a change of trim and rise. The model was tested
at the speeds of: UI = 2.0ms 1, U2 = 3.0ms 1 and U3 =

The forces and moments acting on the hull in the centre

of reference :JOR as a function of speed, drift, trim, roll
and rise have been published in Reference [17]. The
forces and moments acting in any arbitrary centre of
gravity CG are found after transforming the values from
the COR to the CG.

Pitch 6:

4.

MATHEMATICAL MODEL

For planing ships, the pitch varies due to speed variations

and has great effect on the lift and drag and on the
dynamic stability. A combination of low pitch angle and roll
or drift can result in large yaw moments. The model was
62 = l' and 63 = 5'.
tested at the pitch angles of: 6 r =

4.1

FORCES AND MOMENTS BASED ON

EXPERIMENTAL DATA

s.oms',

Using the data obtained from the model experiments in

tabular form in the computer program poses two difficulties. First. interpolating in a five dimensional parameter
space is rather complex, especially since certain combinations of variables were not used during the experiments.
Secondly, because of the six degrees of freedom, the
amount of memory needed during the simulation would be
rather high. It was therefore decided to describe the data
by mathematical formulations derived from regression
I

anarysis.

-e.

Drift

p:

To study the effect of drift, the model was tested at three

.
drift angles of: PI = ir. P2 = 5' and P3 = 1rt.
Roll

a:

Also the influence of the roll angle has been determined

as it aHects the transverse and course stability of the
planing ship. The model was tested at four roll angles of:
9 1 = o; 9 2 = 5'1,6 3 =
and 6 4 = -5'. The negative roll
angle was only tested with Model 277, to study the effect
of symmetry.

The hydrodynamic forces acting on the hull comprise of

linear as well as non-linear components. The basic
structure of the analysis of the mathematical model oi the
hydrodynamic forces is described in full detail in
Reference [16].

ur

Rise of COR

1:

It must be noted that the mathematical model found in this

study is only valid to describe the forces and moments
acting on the two ship models, Model 233 and Model 277,
because the coefficients in the mathematical descriptions
are not presented in non-dimensional form. In further
studies, the scale effects should be examined to be able
to predict stability and manoeuvring characteristics for fullsized planing ships.

The rise of the centre of reference z has great influence

on the behaviour of planing ships. When the ship
accelerates from zero to full speed, first the ship sinks
more into the water, while at higher speed the lift force
pushes the ship out of the water. Assuming the design
draughts of both models were T = O.080m, a negative
rise of z = -5mm and a positive rise of 15mm were
chosen, corresponding to draughts of T = O.085mm and
T = O.065mm respectively.

4.2

DAMPING FORCES

It was desired to run the computer simulation program

without having determined the damping of most of the
motion components. Therefore use has been made of
some rough values of the damping coefficients in the roll
and pitch motions.

The ranges of the variables were chosen after examining

previous results of model experiments with Model 233 [8]
and Model 277 [6].
Some combinations of the variables were skipped during
the experiments due to expected problems with spray.
The total number of test runs was 304.

In future studies additional tests will be conducted to

determine the damping of planing hulls at a higher

accuracy. In this aspect it is thought that only roll and

pitch decay tests will already provide more information
about the damping factors.

Simifar to the formulation of Kp> the damping coefficient

Muq is though: to be implemented in Mq through the speed
dependency In ICe.

For the present research, the damping coefficients in the

Z direction are supposed to be incorporated in
the mathematical model. tt is also assumed that the
couple terms are comparatively small and can therefore
be neglected. The remaining damping coefficients K(p),
M(q) and N(r) remain to be determined in more detail,
especially as a dependency on the speed.

4.2(c)

Yaw damping

x, y and

Inoue f4j and Hooft f2J give an empirical expression for

the damping coefficient of the yaw motion. Based on
these expressions the following equation is used in the
simulation program:
7t

4.2(a)

Ltane

N =-_pL 2 P(0.2:>+0.039_-Q.56_H1 +0.3_ _)

vr
2
T
L
T

Roll damping

For planing ships with deadrise and hard chines, damping

of roll motion is relatively high, because of the immersion
of the planing area at roll angles. Therefore the
non-dimensional damping factor le. defined as:

where Band T are the time-dependent maximum beam

and draught of the ship. In this study, it is assumed that
the non-linear damping term Nr/r/ is small compared to Nur
and can therefore be neglected.
The damping moment for yaw is now:
4.3

is assumed to have a relatively large value. Using the

following equation yields the damping coefficient b, when
le. and the time-dependent I.. , Mpp and c are known, see
also Rutgersson and Ottosson (14J:

Various methods exist to predict the thrust qf the propeller

as a function of the propeller rate of turning.and the ship's
longitudinal speed. Also approximations exist to take into
account the effect of the lateral motions of the propeller.
Often, however, the exact dimensions of the propeller are
not yet determined in the initial design stage.

b=2'let,:ex
1/(1 +Mpp)\'c

Therefore a simplified description of the propeller effect is

used in the simulation program which is assumed to be
acceptable when the propeller RPM are not affected by
the motions of the ship. It should be noted that these
formulae assume that the axis of the propeller shaft is
parallel to the ship-fixed x-axis.

The spring coefficient c is taken from the mathematical

model of the roll moment and depends on the position and
speed of the ship, while the added mass coefficient Mop is
determined below.
In the present study, le. has been varied in the simulation
program in order to ascertain the influence of the roll
damping coefficient. From experimental observations with
free running ships sailing a straight course at high speed
at an initially non-zero roll angle, it was found that the
decay of the roll would occur during a limited number of
oscillations until a stable situation is reached. The value
of le. should therefore be chosen such that also during the
simulation the number of oscillations is found to be small
(approximately one or two). It is expected that 1(. is speed
dependent and may therefore vary in time as a consequence of the change of speed.
The damping moment for roll is now:

4.2(b)

PROPELLER FORCES

For this study one determines the propeller thrust

from:

Xprop

in which Dp is the diameter of the propeller and n the

number of revolutions per second. The thrust coefficient
KT is described by:

Kc~=Kpp=-bp

in which the advance ratio J is defined by: J=

Pitch damping

u '(1-w)
p

-o,

where up is the propeller inflow velocity and wp is the

propeller wake fraction. The coefficients Kn have to be
determined otherwise and are required as input to the
simulation program.

In this study, the pitch damping will be modelled similar to

the roll damping. The following equation is used for
determining the time-dependent pitch damping:

Using the distances between the propeller and eG, YpropYg

and zprop-Zg' the pitch and yaw moment induced by the
propeller thrust are calculated with:

b=2-.:.jU
'<1
!'Y +Mqq )'c
The damping moment for pitch is now: Mcarrp=Mq'q=-b'q

M prop

~ Zprop- Za) Xprop

Nprop=-{yprop-Y

s:

The current.formulae do not take the propeller torque into

account to Induce a roll moment.
4.4

dimensions are not determineo in detail. It is assumed

that the velocities around the rudder are high and that
flow separation does not occur. This last assumption
should be re-evaluated in future studies to increase the
accuracy of the prediction of the rudder forces. For this
preliminary study. added mass and damping of the
rudders are neglected. To increase the accuracy. these
factors are to be included in future studies .

RUDDER FORCES

The formulation of the rUdderforces is based on Inoue [5J

and Hooft [2J. [3]. The rough approximations are suitable
for the preliminary design process. when the adual rudder

...

...

canter
line

r-----------,L---.....::::::..-.:--~o:::__~;:::___-----..,_-

new
Figure 2 Forces acting on the rudder

T~ determine the forces on the rudder. as defined in

Flgu~e 2. the local effective rudder inflow velocity and the

The effective angle of incidence of the flow to the rudder

apparent angle of incidence have to be calculated.

follows from: S =O-OH where 0H=arctan_.

The local rudder inflow velocity for a rudder with average

height h; average chord length c lateral rudder area A
if '
,.
r
and e, sctive aspect ratio A is approximated by:

The lateral rudder force can now be determined, using the

above equations, with:

vr

where
ff"
'" 6.13A. [2] rudd er I'ft
I coe icient
A....2.25

Where

C
Du

o,

0.77i; [3J or 0.9

h: [2] effectiveness of velocity

The lift induced drag in the direction of the rudder inflow

is described by:

increment

where
Cd b

'"

0.7 flow straightening factor

'" 1.0

91

angle between rudder and vertical plane

xr,Yr,zr

position of rudder relative to CG

COl

L&

1tA.

rudder lift induced drag coefficient

The friction resistance of the rudder due to the friction

drag of the rudder is formulated as:

where

ADDED MASS

4.6(a)

Considerations

in this section. the description of the added mass of the

planing ship models will be formulated based on descriptions from strip theory. The symbol used for the added
mass of a strip at position x in direction ifor an acceleration in direction k is:

.. 2-A r rudder wetted area

.. 0.017 (3) high-lift rudder friction coefficient
.. 0.007 [3J NACA profile friction coefficient

mil<(x)

The normal force on the rudder due to the lateral drag

coefficient CN .. 1.8 is:

m;.(x) =0

~(Y,posb-Rsin8-DsinSJx,...(x,+ah,xh)LcosSJcose

.. 0.672Cs-D.153 [2]

mu

m..e

myy

m,..

mY"

jm""

mzz

mza

m,..

mu

jm..e

mza

moo

10

mY"

:"j

increase of rudder efficiency due to induced

force on ship's hull

Imxx
1

where

."

"'''

1n most reports, the added mass of a section with half

beam b is taken proportional to the mass of a semi-circle

C.gx, distance of induced hull force to centre

of gravity
4.5

for i + k odd

With these considerations, the added mass matrix for a

strip at position x looks like:

Krud=- Yn.xJZ,+Z'vdY'
M,vd=X,vd'z, -z.;

..

x,y,z,Q,e,'l'

or

and on grounds of symmetry:

X,ud=-Rcosb-[k;OSb H - Y~inb-LsinbH
Yrud=( -Rsino-DsinoH+ Y,poso+(1 +ah)LcosoH),cOS9 F
Z,vd=( -Rsinb-DsinbH+ Y,posb+(1+a h)Lcosb H ) 'sine F

xh

i,k c 1 ... 6

mik{x)=ml<J{x)

Due to these rudder forces, one finds the following

descriptions for the rudder induced forces on the ship:

with

In Reference [10J, Papanikolaou formulated the added

mass per unit length for sway. heave and roll of a floating
cylinder using potential theory. It can be shown that:

YN=":pA,CNU,..,sino.1 U,..,sino.1
2

N,<Jd=-X"",Y,

4.6

with radius b and specific mass p:

TOTAL EXCITATION FORCES

~ij=~1!b2p.~(~, T, ...).

When the chines of the ship are not immersed, the

draught T of the ship is a measure for the half beam b,

In this preliminary study, the forces and moments dealt

with in the previous sections are supposed to be sufficient
to predict the forces and moments acting on the tested
planing hull forms sailing in calm water. To predict the
behaviour of a vessel in all weather and sea conditions,
descriptions to model the influence of waves, wind and
current on the ship have to be incorporated.

therefore one can also write:

mij=~1!T2p'f2(~' T, ...).

This factor can also be seen in the formulations for the

added mass in this paper.
4.6(b)

It is generally known that the wave forces have a large

influence on the behaviour of the ship, but the influence of
wind can also be large. A strong side wind can induce
large roll angles, changing the hydrodynamic forces and
moments considerably. Strong wind gusts can result in
coupled roll-yew motions, possibly resulting in broaching
or capsizing.

Added mass for x-direction

Because the values of mu and m..e are presumably small

compared to mu' these added masses are taken to be
zero in the present study. In future studies, these
components can be assigned non-zero values if desired
to increase accuracy.
For this study, the total added mass
mated by

The total excitation forces and moments about the centre

of reference in ship-fixed directions are found by adding
all force components:

where Cm is taken as: Cmz

Mu can be approxi-

= 0.8. The parameters Tma:c

and B/Tlilx are the time-dependent instantaneous maximum
draught and breadth at the still water line of the ship.
%

This equation is used for calculating the accelerations of

the ship.

4.6(c)

4.6(d)

Added mass for v-direction

Added mass for zdirection

The added mass for an acceleration in z-direction is

described by Payne (11) and Quadvlieg [13]. Both reports
give the following description for the added mass per unit
length for a section with deadrise angle p and draught T
as:

Papanikolaou gives in his report tables to determine the

added mass in y-direction for variable excitation
frequencies and breadth to draught ratios. In the present
study, the values for Cl) lE 0 S1 should be used, because
the equations of motions are solved for a quasi-static
state of the ship.
Diagram 4 from [10] gives the non-dimensional added
of a section as a function of the
mass coefficient
beam to draught ratio. The values in this figure for zero
oscillation frequency will be approximated in this study
using the following function:

mrr'

The function f(13) gives the quotient of the added mass for
a prism and the added mass for a flat plate. According to

p in radians,

Payne, f(l3), with

Quadvlieg gives:

f(P)=1-~.

is given by

f(~)=1 -i

while

7t

For zero deadrise, a flat

7t

plate, both deadrise functions yield the same value, Le.

The coefficients myy; are determined by performing
regression analysis. This results in the following
coefficients:

frO) = 1. However, for a dead rise of

2:,
2

the function by

Payne gives f('::)=.2.. while the function of Quadvlieg

2 2

results in f(-i')=O.

hi9J~]

Care should be taken using this formula lor

In this study, it is decided to use the description by

Quadvlieg in the simulation program. Further study should
examine the added mass in more detail to determine
which formulation yields more realistic results. Recent
work by Payne [12] gives suggestions on improvements
in the determination of the added mass.

values (greater than approximately 5). Because of\he

regression model, e'Xtrapolating will yield unrealistic
results.
The added mass in y-direction per unit length for a section
with draught T can now be derived with:

This added mass per unit length mzz can now be used to
calculate the foilcwing added masses:

MzG=MezJL
=( m -x dx
zz

When myy is known, the following added masses can be

calculated:

=(m

""" JL

yy

4.6(e)

x2 dx

Added mass for a-direction

The hydrodynamic mass in c-olrectlcn, depending on the

beam to draught ratio and the oscillation frequency is
given in Diagram 7 of the report by Papanikolaou [10].
Again, the values for zero oscillating frequency should be
used. Similar to myy', these values will be approximated in
this study using the following function:

Papanikolaou also gave a relation between the added

masses m~ and myy by plotting the virtual arm h.

M66 =(
m x2 dx
JL zz

m
=_Y_..
mr-T
yy

These values will be approximated using mj1 =myy' T'hj1

and

The coefficients m..i are determined by regression

analysis, resulting in:
Performing regression analysis
coefficients for the virtual arm:

hj10=O .4472

hr.' =-0.0218

yields the following

hr+2=-o13!\-9

hj13=-o0007

The added mass Mj1 can be calculated with:

Note again that extrapolation will probably result in

unrealistic values. The added mass in o-dlrectlon per unit
length for a section with draught T can now be calculated
with:

5.

B
1tpr_

m.. =

TmI

In this section, the simulations performed during the study

are described. The simulations were done to ascertain
whether the program is working properly to evaluate the
results of the computations. First the set-up of the test
program is discussed, after which the results of the
simulation are given and discussed. Details about the
computer program can be found in Reference [18].

Using m.. , the added mass for roll is found by:

5.1
4.7

EQUATIONS OF MOTIONS

1.

Change in model set-up, e.g. change of mass or

position of CG.

2.

Change in initial position or velocity, e.g. change of

rise or speed.

3.

Change in hydrodynamic coefficients, .i.e. damping

factors and deviation from equilibrium position.

4.

Change in manoeuvring mode, e.g. turning Circle test

and zig-zag test.

second law: M" J:.r=~:

'x -

m(qw-rv)

m(ru-pw)

m(pv-qu)

Iy
Iz

K ..

M"'lM

1r:

(fyy-Iu)qr

.. 'x?q

- '.IP'-r)
- ':cqr

.. (lzz-/u)pr

rr P I'

N .. (lu-1yy)pq

The first type is used to determine the trends in the

results due to changes in the input. The outcome of some
changes can be predicted using general theory or
publications about this subject.

in which the mass matrix M is:

m..Myy

My<

M,...,

m..Mzz

Mze

M~

'u+ M..

-I +M

Mze

Iyy .. Moo

Myy

-1:a..M.v

Izz+M"",

xz

The second type is used to determine the ability of the

ship to return to its equilibrium position, irrespective of the
initial deviation from this equilibrium.

m..Mxx

M =

The third type is used for stability crheria, concerning the

values of these coefficients.

."

The fourth type is used to determine the manoeuvrability

of the planing ship and the behaviour of the ship during
these manoeuvres. Combining the third and fourth types
can yield more strict limits than those found using the
simulations from type three.
Before starting the test program, a few runs were done
with both ship models to determine the values of the
damping factors K. and K'a. A value of K e = 0.6 was found
to be satisfactorily for both Model 233 and Model 277.
The value of K. was for Model 233: K. = 0.6 and for Model
277: K. = 0.55. It appears that Model 277 is less sensitive
to roll velocities than Model 233. This will be examined in
more detail further on in this paper.

Solving the above equation at each time step will give the
accelerations. Integration of these accelerations provides
the velocities u, v, W, p, q and t, which determine the
excitation forces J:.r at the next time step.
Transforming the velocities from the ship-fixed coordinate
system to the earth-fixed system and subsequently
integrating these earth-fixed velocities will provide the
position of the ship in the earth-fixed system.

Weight

TEST PROGRAM

The test program consists of various types of simulation

runs. The following types can be defined:

To calculate the ship-fixed accelerations ~ of the planing

ship, the equations of motions are derived from Newton's

SIMULATIONS

The first run for both models is derived from model

experiments performed by Keuning [8J, (Model 232-A is
Model 233 in this study), and [7], (Model 188 is Model 277
in this study). The following conditions were selected:

164 N

2.3 ms-t

3.7'

8S/m
3.7'

-4.5 mm

-3.1 mm

Model 277

159.9 N

2.4 ms-1

2.3'

1.5'

-5.3 mm

-3.9 mm

Rise z'XP

Rise

Model 233

zs:m

It appears that the mathematical model describes the

state of Model 233 rather satisfactorily. For Model 277,
the similarity between the test result and the simulation
result is less clear. The difference may be caused by
differences between the model test set-up and the
mathematical model. Such differences exist for example
in the set-up of the propulsion and the modeiling of the
rudder forces.

5.2

SIMULATION RESULTS

In this section the results of some simulation runs will be

presented. For the set of runs where the model set-up
was changed, the results stated in Table 1 were found.

TABLE 1 Results of first set of simulations

Run ID

Description

Expected behaviour

Simulation result

R233-12

Shift LCG aft

R23313

Shift LCG forward

e
Decrease e

R233-14

Decrease GM

Increase 8

R233-15

Increase GM

Decrease 8

e = 3.72' ~ e =4.3d'
e = 3.72' ~ e =3.22'
8 =3.72' ~ e = 3.73'
e = 3.72' ~ e = 3.71

R233-16

Decrease mass

Decrease draught

T = 0.088m ~ T

R233-17

Increase mass

Increase draught

R277-12

Shift LCG aft

Increase 8

= 0.088m ~ T = 0.093m
e = 1.4g" ~ e = 2.02'

R277-13

Shift LCG forward

Decrease

R27714

Decrease GM

R277-15

Increase GM

e
Increase e
Decrease e

e = 1.4g" --7 e = 1.02'

e = 1Ag" ~ e = 1.5d'
e = 1Ag" ~ e = 1.4g"

R277-16

Decrease mass

Decrease draught

R27717

Increase mass

Increase draught

Increase

= 0.082m

= O.085m ~ T = 0.079m
T = 0.085m ~ z = O.OgOm

In Figure 4 the results of the vertical position of the centre

of gravity during runs R233-23 and R277-23 are
presented in graphical form. Clearly, the oscillations due
to the zero initial vertical (shipfixed) velocity are larger for
Model 233 than for Model 277. However, the vertical
velocity in the equilibrium state is also much higher for
Model 233 than for Model 277. Examining the results, it is
seen that the vertical oscillations are a result of the
combined heave-pitch system.

For all simulations, it appears that the behaviour of the

computer program to changes in model set-up is
consistent with theory or experience. Also, both ship
models respond similarly to the changes in model set-up.
However, although the damping factor Kg is larger for
Model 233 than:for Model 277, the oscillations around the
equilibrium after the initial disturbance are more
pronouncedfor Model 233.

In the graphs taken from the results of runs R233-13 and

R277-13, see Figure 3, the difference in oscillation
amplitude and duration can be seen. This can be
explained by examining the mathematical model for the
pitch moment M, see Reference [15Jfor more details. The
spring t~rm to calculate the damping moment is much
larger for Model 277 than for Model 233. For Model 277,
any pitch angular velocity will be damped stronger in
comparison to Model 233. Because the damping is still
unknown it is not possible to draw any conclusions on the
difference in behaviour between the to hull forms.

In Table 3, the results of the third set of runs are listed. It

is clearly seen that decreasing the damping factors largely
affects the stability of the ship. The limit values of the
damping factors le. and Kg below which the behaviour of
the ship becomes unstable are probably as follows:
Model 233:

0.15 <

Model 277:

0.1375 <

K.

< 0.30,
K.

0.15 <

< 0.275,

K9

< 0.30

Kg<0.15

In future studies, the limits for the damping factors can be

determined more accurately.

Table 2 was constructed after simulation of the runs, with

deviations from the initial equilibrium state. At the end of
each simulation, the state of the ship model was returned
to the equilibrium state, Le. sailing at straight course and
constant speed. It can therefore be concluded that for the
used input the state of the planing vessel is stable. From
runs R233-24 and R277-24, it also appears that the
vessel is course stable.

As an example of the behaviour of the model at low

damping factors, the results of runs R233-34 and R233-36
are included in Figure 6. The unstable behaviour when the
damping factor for roH is decreased is clearly visible in the
increasing oscillation amplitudes of the roll angle and the
increasing immersion.

TABLE 2 Results of second set of simulations

Run ID

Initial deviation

Simulation result

R23321

Decrease of speed

R233-22

Increase of speed

R23323

Zero vertical speed

R233-24
R233-26

= -S'
T - 5 mm
Draught T + 5 mm

R27721

Decrease of speed

Ut.o = 1.285 ms" -t u,_ = 2.285 ms"

u,-a '"' 3.285 ms" -t u,_ '"' 2.285 ms"
w,.o = 0.000 ms" -t w,_ = 0.149 ms"
v,.o = 0.200 ms" -t v,_ = 0.000 rns"
T,.o '"' 0.0827 m -t T,_ = 0.0877 m
T,.o =0.0927 m -t T,_ = 0.0877 m
U,.o = 1.370 ms" -t u,_ '"' 2.370 rns"
ut-a =3.370 ms" -t u,_ = 2.370 ms"
w,.o '"' 0.000 ms' -t wt_ = 0.062 rns"
vt-a '"' 0.207 ms'! -t vt_ = 0.000 rns"
T,.o = 0.0796 m -t T,_ = 0.0846 m
T,-a = 0.0896 m -t T,_ = 0.0846 m

R23325

Initial drift ~
Draught

R27722

Increase of speed

R277-23

Zero vertical speed

R277-24

Initial drift ~

R277-25

Draught

R277-26

Draught

= -S'

T - 5 mm
T + 5 mm

TABLE 3

Results of third set of simulations

Run ID

Description

Simulation result

R233-31
R23332

Zero pitch angie, Ke = 0.60

lCe = 0.5 1(e,o = 0.30

Decreasing oscillations

R233-33

lCe = 0.25 lCe,o = O. 15

Increasing oscillations

R233-34

Initial roll angle Cl = 5,

R23335

K. =

R233-36

K.

= 0.25 K.,o = 0.15

Increasing oscillations

R23337

Roll angle S', zero pitch

Decreasing oscillations

R23338

K.

= 0.5 K.o' lea = 0.5 lea,o

Zero pitch angle, K e = 0.60

Decreasing oscillations

Ke = 0.5 Ke,o = 0,30

Ko = 0.25 ;;,0 = 0.15

Decreasing oscillations

R277-33
R27734

Initial roll angle Cl

R277-31
R277-32

0.5

Ko

K.

Decreasing oscillations

= 0.60

Decreasing oscillations
Decreasing oscillations

= 0.30

Decreasing oscillations
Decreasing oscillations

= 5, K. = 0.55

R27735

K.

= 0.5 K.,o

R277-36

K.

= 0.25 K.,o = O. 1375

R277-37

Rol! angle 5', zero pitch

R277-38

K.

Decreasing oscillations
Decreasing oscillations

= 0.275

= 0,5 K.,o'

lea

Increasing oscillations
Decreasing oscillations

= 0.5 lea,o

Decreasing oscillations

TABLE 4 Results of turning circle tests

Run ID

Tactical

Advance

Transfer

Diameter
R233-41

11.37

13.50

5.77

Damping

Approach

Speed

Speed

Loss

ms"

K.

lCe

2.29

7.6

0.60

0.60

Roll

Pitch

R233-42

11.25

13.62

5.76

2.29

8.5

0.30

0.30

R277-41

12.68

14.45

6.47

2.37

7.1

0.55

0.60

2.37

7.3

0.27

0.30

+12%

+50%

50%

+50%

50%

R277-42

12.95

14.66

6.62

233: Difference

-1.1%

+0.9%

0.2%

277: Difference

+2.1%

+1.5%

+2.3%

R233-TT

11.37

13.50

5.77

2.29

7.6

0.60

0.60

R277-TT

12.78

14.31

6.48

2.28

7.0

0.55

0.60

+12.4%

+6.0%

-0.4%

-9.2%

-8.3%

+0%

Difference

+12.3%

10

+2.8%

The simulation results of the turning circle tests for both

ship models are summarized in Table 4. The first run of
each ship model is with the original damping factor values
and with a rudder angle of 35 to starboard. During the
second run, the damping factors were divided by two. It
should be noted that, except for runs R233n and R277
Tl, the approach speed of both models is not exactly the
same (difference approximately 3.5%), therefore the
results are not similar. In these runs, a small influence of
the damping factors on the manoeuvring characteristics is
seen.

the calculation results. Some of these

discussed in this paper.

Run Id

7.

Roll

First
overshoot

Second
overshoot

Pitch

Angle

Angle

R233-43

0.60

0.6

27.3

31.5

R233-44

0.30

0.3

28.9

34.3

R277-43

0.55

0.6

26.5

30.4

R277-44

0.275

0.3

27.1

31.2

233: Difference

-50% -50%

+5.9%

+8.9%

277: Differen~e

-50% -50%

+2.3%

+2.6%

Also thanks to Prof. Dr. Ir. J A Pinkster at Delft University

of Technology for his support and guidance during this
project. The authors also wish to thank the members of
the Ship Hydromechanics Laboratory for their help during
preparation and performing of the model experiments.

8.

The results from the 20/20 zig-zag tests are stated in

Table 5. In this table it is seen that for this manoeuvre
Model 233 reacts stronger to changes in the damping
factors than Model 277. It appears that decreasing the
damping results in an increase of the overshoot angles.
In Figyre 8 the time samples of both the rudder and yaw
angles during run R27743 are included. Note that a
starboard rudder angle has a positive sign in this figure.

6.

ACKNOWLEDGEMENTS

The authors wish to express their gratitude to the Maritime

Research Institute the Netherlands in Wageningen for
supporting this research.

Results of 20/20 zig-zag tests

Damping

are

Examining the simulations, it is seen that changes in the

input of the program resulted in the expected changes in
the output. Changing the damping factors appears to have
great influence of the behaviour of the ships. Further
study should emphasize in determining the dynamic
coefficients in the mathematical model more accurately.
In general it is concluded that the results from this
computer simulation program can be used in the early
design stage to predict the stability and manoeuvrability of
the planing ship.

Runs R233n and R277n were performed to illustrate

the difference in manoeuvrability between both ship
models. The mass, position of centre of gravity and the
approach speed were identical for both ships. In Table 4
and Figure 7 the results of the simulations are included.
During the simulations, speed losses of approximately 8%
were found, while in reality the speed loss during close
turning of planing ships can reach values of 70% or even
higher. Implementing rotational velocity dependent
hydrodynamic coefficients such as XV!" in the mathematical
model in the future should increase the accuracy of the
simulation program considerably.

TABLE 5

results

1.

CLEMENT, E P and BLOUNT, 0 L: 'Resistance

tests of a systematic series of planing hull forms'.
SNAME Transactions, Vol. 71, 1963.

2.

HOOFT, J P: Computer simulation of the ship's

manoeuvrability, Part 1 & 2. Maritime Research
Institute the Netherlands.

3.

HOOFT, J P and NIENHUIS, U: 'The prediction of

the ship's manoeuvrability in the design stage'.
SNAME Annual Meeting, November 1994.

4.

INOUE, S, HIRANO, M, and KIJIMA, K:

'Hydrodynamic derivatives on ship manoeuvring'.
International Shipbuilding Progress, Vol. 28, No. 321,
pp. 112-125, 1981.

5.

INOUE, S, HIRANO, M, KIJIMA, K, and TAKASHINA,

J:
'A practical calculation method of ship
maneuvering motion'.
International Shipbuilding
Progress. Vol. 28, No. 325, pp. 207-222. 1981.

6.

KEUNING, J A:
Invloed van de Deadrise op het
zeegangsgedrag van planerende schepen. Report
794-0 Delft University of Technology, June 1988.

7.

KEUNING, J A and GERRITSMA, J: 'Resistance

tests of a series of planing hull forms with 25 degrees
deadrise angle'. International Shipbuilding Progress,
Vol. 29, No. 337, pp. 222-249, 1982.

CONCLUSION

In this paper, a time-domain computer simulation program

to predict the dynamic stability and manoeuvrability of
planing ships in still water for six degrees of freedom is
described. The formulations used in the program were
based on experimental data and additional empirical
coefficients taken from literature.
Simulation runs have been performed to ascertain
whether the program is working properly and to evaluate

11

REFERENCES

8. KEUNING, J A, GERRITSMA, J, and VAN

TERWISGA, P F: Resistance tests of a series
planing hull forms with 30 degrees deadrise angle,
and a calculation model based on this and similar
........
t....... cn~~lL
...............
~'Il:,.n:J~.

n~n
o:1iJwl

15. TOXOPEUS, S L: A time domain simulation program

for manoeuvring of planing ships. Deltt University of
Technology, September 1996.

r\,..U+ II ....h ,,.. .... :..., ..... " T ....""'I-. ......... I..... ,.u,

""t::JII\

VIIlYc::JI~1L1

\,11

16. TOXOPEUS, S L:

Mathematical mode! of the

behaviour of planing ships. Delft University of
Technology, August 1996.

1c::J....,IIIIVIV~1,

December 1992.
9. LEWANDOWSKI, E M: 'Trajectory predictions for
high speed planing craft'. International Shipbuilding
Progress, Vol. 41, No. 426, pp. 137148, 1994.

17. TOXOPEUS, S L: Model experiments on dynamic

stability of planing ships. Delft University of
Technology, June 1996.

10. PAPANIKOLAOU, A: 'Hydrodynamische Koeffjzienten

fOr die linearen Schwingungen von schwimmenden
Zylindern'.
Schiffstechnik, Vol. 27, No. 3, pp.
127166,1980.

18. TOXOPEUS, S L: VesSim User's Manual. Maritime

Research Institute the Netherlands, to be published.
19. ZARNICK, E E: A nonlinear mathematical model of
motions of a planing boat in regular waves. Report
DTNSRDC78/032, David W. Taylor Naval :Ship
Research and Development Center, March 1978.

11. PAYNE, P R: 'The vertical impact of a wedge on a

fluid'. Ocean Engineering, VoL 8, No. 4, pp. 421-436,
1981.
12. PAYNE, P R: 'Recent developments in 'added-mass'
planing theory'. Ocean Engineering, Vol. 21, No. 3,
pp. 257309, 1994.

20. COHEN, S H, BLOUNT, D L: 'Research Plan fer the

investigation of Dynamic Instability of ,small High
Speed Craft', SNAME Transactions, Vol. 94, 1986, pp
197214.

13. QUADVLlEG, F F H A: Non linear motions of planing

ships. Report 920S, Delft University of Technology,
March 1992.

21. BLOUNT, D Land CODEGA, L T: 'Dynamic Stability

ef Planing Boats', Marine Technology, Vol. 29, No. 1,
January 1992, pp 412.

14. RUTGERSSON, 0 and OnOSSON, P: 'Model tests

and ccmputer simulations an effective combination
for investigation of broaching phenomena'. SNAME
Annual Meeting, New York, N.Y., November 1987.

VesSim Sample Recording

Pitch angle

c;

:E.3

f\.

"'-"'-"'~--------------------------------

, .~~---------------------------o .

1-

Fig.3

8
iim.=.!c:l

R23313 -

10

R27713

12

Results of simulation, runs R23313 and R27713

12

14

16

NOMENCLATURE
Symbol

Description

Unit

CB
GM
K
L
LCG
LCOR
M
N
p
q
r
T
u
ue
U
v
ve
VCG
VCOR
w
X
xe
xG
xR
Y
ye
yG
yR
z
Z
ze
zG
zR
b
q
f

Design block coefficient

Metacentric height
Ship-fixed moment in longitudinal direction
Length between perpendiculars
Longitudinal position centre of gravity
Longitudinal position centre of reference
Ship-fixed moment in lateral direction
Ship-fixed moment in vertical direction
Rate of turning around x-axis
Rate of turning around y-axis
Rate of turning around z-axis
Draught of the model at centre of reference
Ship's longitudinal velocity
Ship's longitudinal velocity, earth-fixed
Towing speed
Ship's lateral velocity
Ship's lateral velocity, earth-fixed
Vertical position centre of gravity
Vertical position centre of reference
Ship's vertical velocity
Ship-fixed force in longitudinal direction
x-position of centre of gravity, earth-fixed
x-position of centre of gravity, ship-fixed
x-position of centre of reference, ship-fixed
Ship-fixed force in transversal direction
y-position of centre of gravity, earth-fixed
y-position of centre of gravity, ship-fixed
y-position of centre of reference, ship-fixed
Rise of centre of reference, coordinate along z-axis
Ship-fixed force in vertical direction
z-position of centre of gravity, earth-fixed
z-position of centre of gravity, ship-fixed
z-position of centre of reference, ship-fixed

m
Nm
m
m
m
Nm
Nm
rads-1
rads-1
rads-1
m
ms-1
ms-1
ms-1
ms-1
ms-1
m
m
ms-1
N
m
m
m
N
m
m
m
m
N
m
m
m

Drift angle, twist around earth-fixed z-axis or deadrise angle

Trim angle, twist around ship-fixed y-axis

deg
deg

Roll angle, twist around ship-fixed x-axis

Yaw angle
Rudder angle, positive to port

deg
deg
rad

Desired rudder angle

Damping factor for roll damping
Damping factor for pitch damping

rad
-

y
d
dd
kf
kq

13

VesSim Sample Recording

z-pos
0.016
0.014
0.012

z-pos [m]

0.010
0.008
0.006
0.004
0.002
0.000
-0.002
-0.004
0

8
Time [s]
R233-23

Fig. 4

10

12

14

16

14

16

R277-23

Results of simulation, runs R233-23 and R277-23

VesSim Sample Recording

Roll angle
5
4
3

phi [deg]

2
1
0
-1
-2
-3
-4
-5
0

8
Time [s]
R233-24

Fig. 5

10

12

R277-24

Results of simulation, runs R233-24 and R277-24

14

VesSim Sample Recording

Roll angle
6
5
4
3

phi [deg]

2
1
0
-1
-2
-3
-4
-5
-6
0

10
Time [s]
= 0.60

Fig. 6

15

20

= 0.15

Results of simulation, runs R233-34 and R277-36

VesSim: 20/20 Zig-zag test

Yaw and rudder angles
60

Yaw and rudder angles [deg]

40

20

-20

-40

-60
0

10

20

30
Time [s]

R277-43: Yaw

Fig. 8

R277-43: Rudder

Results of simulation, run R277-43

15

40

50

VesSim: Turning test

15
35
12

30
20

15

15

9
30
20
6

35
40
25
10
10

y [m]

25
0

42

-3
44
40

-6

-9
42
R233-TT

-12

R277-TT

-15
0

Fig. 7

12

15
x [m]

18

21

24

Results of simulation, runs R233-TT and R277-TT

16

27

30